Skip to main content
Use the options below to give AI coding assistants and language models access to MoEngage documentation with less overhead than loading full HTML pages.

​
Quick start

  • AI coding assistants (Claude Code, Cursor, etc): Use the MoEngage Docs MCP server for the best experience. It lets your assistant search and fetch documentation from within the IDE.
  • Chat interfaces (ChatGPT, Claude.ai): Add .md to any docs page URL and share it in your conversation to get plain markdown instead of the full site.
  • Project rules and custom instructions: Add MoEngage’s llms.txt URL to your AI tool’s rules so it has an index of relevant MoEngage pages.
  • Agent capabilities: Use the skill.md URL or npx skills add https://www.moengage.com/docs so agents know what they can do with MoEngage and what inputs or constraints apply.

​
MCP server

The MoEngage Docs MCP server gives AI agents and coding assistants direct access to MoEngage documentation and references through the Model Context Protocol. Use it to bring MoEngage knowledge into tools like Claude, Cursor, and other MCP-compatible clients. The MoEngage Docs MCP (Model Context Protocol) server connects your AI agents to MoEngage’s docs and developer resources. Configure the MoEngage Docs MCP server in your IDE or AI chat to get in-context help for your MoEngage integration.
For more on the Model Context Protocol, see the MCP documentation.
If your AI agent supports MCP, you can add the MoEngage Docs MCP server so it has access to MoEngage’s APIs and SDKs. The MoEngage Docs MCP server lets you:
  • Get quick answers about MoEngage APIs, features, and implementation
  • Search documentation without leaving your editor
  • Find code samples and recommended patterns for typical use cases
  • Debug issues using docs and SDK references
  • Write code that integrates your app with MoEngage
  • Report incorrect or outdated documentation to the MoEngage docs team

​
Installation

To set up the MoEngage Docs MCP server, use the contextual menu on any page and click Connect.
Cursor Context 1
Alternatively, you can pick your IDE or AI chat provider and follow the steps below.
  1. Open Cursor and open the .cursor directory in your project root. Create it if it doesn’t exist. This is where Cursor’s project configuration lives.
  2. Open the mcp.json file in .cursor. Create it if it doesn’t exist. This file holds MCP server configuration for the project.
  3. Add the following to mcp.json:
    mcp.json
  4. Save the file.
  5. Restart Cursor.
Claude Desktop uses npm to run the MoEngage Docs MCP server. You need Node.js installed. See Node.js downloads for details.
  1. Open Claude Desktop.
  2. Go to File > Settings (Windows) or Settings (macOS).
  3. Open the Developer tab.
  4. Click Edit Config.
  5. Open claude_desktop_config.json. This file controls MCP servers for Claude Desktop.
  6. Add the following to claude_desktop_config.json:
    claude_desktop_config.json
  7. Save the file.
  8. Restart Claude Desktop.
Remote MCP in Claude Desktop is currently in beta.

​
Enterprise and team plans

  1. Open Claude Desktop.
  2. Go to Admin settings > Connectors.
  3. Scroll to the bottom of the section.
  4. Click Add custom connector.
  5. Fill in:
    • Name: MoEngage Docs MCP
    • Remote MCP server URL: https://www.moengage.com/docs/mcp
  6. Click Add.

​
Pro and max plans

  1. Open Claude Desktop.
  2. Go to Settings > Connectors.
  3. Scroll to the bottom of the section.
  4. Click Add custom connector.
  5. Fill in:
    • Name: MoEngage Docs MCP
    • Remote MCP server URL: https://www.moengage.com/docs/mcp
  6. Click Add.
Run the following command in your terminal:
  1. Create a .vscode/mcp.json file with the following content:
    mcp.json
  2. Save the file.
  3. Restart VS Code.

​
Usage (Tools)

The MoEngage Docs MCP exposes the following tools:
  • search_mo_engage — Searches MoEngage’s documentation and knowledge base. Use it for broad or conceptual queries such as “how to authenticate” or “rate limiting”. Results include contextual content with titles and links to the relevant docs.
  • query_docs_filesystem_mo_engage — Runs read-only, shell-like commands (rg, grep, find, tree, ls, cat, head, jq, and other text utilities) against a virtualized filesystem that contains the MoEngage documentation pages and OpenAPI specs. Use it for exact keyword or regex matches, exploring the docs structure, or reading the full content of a page by its path. Nothing runs on your machine or network — the filesystem is an in-memory sandbox backed by documentation content.
  • submit_feedback — Reports a problem with a documentation page to the MoEngage docs team. Use it when a page is incorrect, outdated, confusing, or incomplete, or when an example is broken.
A typical agent workflow starts with search_mo_engage for conceptual questions, then uses query_docs_filesystem_mo_engage to read the full page (append .mdx to the path returned by search) or to run precise regex searches.

​
Example Prompts

After you connect the MCP server, ask your AI assistant questions in plain language. The assistant searches MoEngage documentation and answers with citations, so you don’t have to switch to the docs site.
Use a chat client such as Claude Desktop to get answers about campaigns, segments, and analytics:
  • “How do I set up a push campaign with send time optimization?”
  • “What is the difference between event-triggered campaigns and Flows? When should I use each?”
  • “How do I create a segment of users who purchased in the last 30 days but haven’t opened the app since?”
  • “How are conversion goals attributed in campaign reports?”
  • “Walk me through setting up a WhatsApp campaign, including the prerequisites.”

​
Per-page markdown

Append .md to any docs page URL to get a markdown version of that page. You get content only, without HTML, nav, or scripts, which uses fewer tokens than the full page. For example, for the API Reference introduction in markdown, use:

​
Site-wide endpoints

These endpoints follow the llms.txt standard. You can use these endpoints for markdown-formatted content.

​
skill.md

The skill.md file describes what AI agents can do with MoEngage in a structured, machine-readable format. It makes capabilities, required inputs, and constraints explicit so agents can use your product more reliably. URL: https://www.moengage.com/docs/skill.md The file follows the skill.md specification and is kept in sync with the documentation so agents stay aligned with current MoEngage features and APIs.
llms.txt and skill.md both help agents use MoEngage documentation, but they do different things.
  • llms.txt is a directory: it lists documentation pages with descriptions so agents know where to look.
  • skill.md is a capability summary: it describes what agents can accomplish with your product, what inputs they need, and what constraints apply.

​
Using skill.md with agents

Agents can load MoEngage’s capabilities into their context with the skills CLI:
This adds MoEngage’s skill description to the agent so it can take actions on behalf of users (e.g. answering integration questions or suggesting API usage).
Use the skill.md URL or the command above in your own AI setup so you get better, more accurate assistance when working with MoEngage.

​
What’s in skill.md

The MoEngage skill.md file is generated from the documentation and typically includes:
  • Metadata — Product name, description, and version.
  • Capabilities — What agents can do with MoEngage (e.g. campaigns, segments, SDK integration).
  • Skills — Concrete actions grouped by area (APIs, SDKs, data, etc.).
  • Workflows — Step-by-step procedures for common tasks.
  • Integration — Supported tools and services.
  • Context — High-level overview of MoEngage documentation and concepts.
For the exact schema and fields, see the agentskills.io specification.